The modified Glasgow Prognostic Score (mGPS) can guide decisions for immunotherapy treatment beyond progression.

Background: Treatment beyond progression (TBP) is common in patients treated with immune-checkpoint inhibitors (ICI), however, there is no biomarker to select patients that are more likely to derive benefit from TBP. Here, we investigated the potential of the modified Glasgow Prognostic Score (mGPS) as a predictive biomarker to select patients for TBP.

Methods: We performed a post-hoc analysis of the immunotherapy arms in the randomized phase 3 trials IMmotion151 (renal cell carcinoma), OAK (non-small cell lung cancer) and IMvigor211 (urothelial cancer). The main outcome was post-progression overall survival (PPOS) after the first investigator-assessed disease progression (PD), in mGPS risk groups. The mGPS classifies patients into three risk groups based on C-reactive protein (CRP) and albumin.

Results: We found a strong prognostic value for the mGPS when assessed at the time of PD (PD-mGPS) in all three trials. High-risk PD-mGPS was associated with significantly shorter PPOS compared to low-risk PD-mGPS (HR for death 18.3 (95 % CI 6.71-50.0, p < 0.001)) for RCC, UC: HR 4.16 (95 % CI 2.58-6.69, p < 0.001) and NSCLC HR 2.53 (95 % CI 1.70-3.77, p < 0.001). Importantly, patients within all three trials only derived benefit from ICI-TBP compared to switch to further-line treatment in the PD-mGPS low-risk group (RCC: HR 0.18 (95 % CI 0.06-0.55, p = 0.002); UC: HR 0.59 (95 % CI 0.34-1.00, p = 0.052); NSCLC: 0.62 (0.41-0.92, p = 0.018) compared to PD-mGPS intermediate/ high risk).

Conclusions: These findings suggest that the mGPS measured at the time of radiologic PD can identify patients with a better prognosis who may benefit from continued atezolizumab therapy, aiding in the selection for TBP in clinical practice.